Medium-term forecast of daily passenger volume of high speed railway based on DLP-WNNMedium-term forecast of dailypassenger volume of high speedrailway based on DLP-WNN

Purpose – This paper aims to propose a medium-term forecast model for the daily passenger volume of HighSpeed Railway (HSR) systems to predict the daily the Origin-Destination (OD) daily volume formultiple consecutivedays (e.g. 120 days).Design/methodology/approach – By analyzing the characteristics of the historical data on daily passengervolume of HSR systems, the date and holiday labels were designed with determined value ranges.In accordance to the autoregressive characteristics of the daily passenger volume of HSR, the Double LayerParallel Wavelet Neural Network (DLP-WNN) model suitable for the medium-term (about 120 d) forecast of thedaily passenger volume of HSR was established. The DLP-WNN model obtains the daily forecast result byweighed summation of the daily output values of the two subnets. Subnet 1 reflects the overall trend of dailypassenger volumes in the recent period, and subnet 2 the daily fluctuation of the daily passenger volume toensure the accuracy of medium-term forecast.Findings – According to the example application, in which the DLP-WNN modelwas used for the medium-termforecast of the daily passenger volumes for 120 days for typical O-D pairs at 4 different distances, the averageabsolute percentage error is 7%-12%, obviously lower than the results measured by the Back Propagation (BP)neural network, the ELM (extreme learning machine), the ELMAN neural network, the GRNN (generalizedregression neural network) and the VMD-GA-BP. The DLP-WNN model was verified to be suitable for themedium-term forecast of the daily passenger volume of HSR.Originality/value – This study proposed a Double Layer Parallel structure forecast model for medium-termdaily passenger volume (about 120 days) of HSR systems by using the date and holiday labels and WaveletNeural Network. The predict results are important input data for supporting the line planning, scheduling andother decisions in operation and management in HSR systems.

Revenue management method and critical techniques of railway passenger transport

Purpose–Revenue management(RM)is a significant technique to improve revenue with limited resources.With the macro environment of dramatically increasing transit capacity and rapid railway transport development in China,it is necessary to involve the theory of RM into the operation and decision of railway passenger transport.Design/methodology/approach–This paper proposes the theory and framework of generalized RM of railway passenger transport(RMRPT),and the thoughts and methods of the main techniques in RMRPT,involving demand forecasting,line planning,inventory control,pricing strategies and information systems,are all studied and elaborated.The involved methods and techniques provide a sequential process to help with the decision-making for each stage of RMRPT.The corresponding techniques are integrated into the information system to support practical businesses in railway passenger transport.Findings–The combination of the whole techniques devotes to railway benefit improvement and transit resource utilization and has been applied into the practical operation and organization of railway passenger transport.Originality/value–The development of RMRPT would provide theoretical and technical support for the improvement of service quality as well as railway benefits and efficiency.

Railway Passenger Flow Forecasting by Integrating Passenger Flow Relationship and Spatiotemporal Similarity

Railway passenger flow forecasting can help to develop sensible railway schedules,make full use of railway resources,and meet the travel demand of passengers.The structure of passenger flow in railway networks and the spatiotemporal relationship of passenger flow among stations are two distinctive features of railway passenger flow.Most of the previous studies used only a single feature for prediction and lacked correlations,resulting in suboptimal performance.To address the above-mentioned problem,we proposed the railway passenger flow prediction model called Flow-Similarity Attention Graph Convolutional Network(F-SAGCN).First,we constructed the passenger flow relations graph(RG)based on the Origin-Destination(OD).Second,the Passenger Flow Fluctuation Similarity(PFFS)algorithm is used to measure the similarity of passenger flow between stations,which helps construct the spatiotemporal similarity graph(SG).Then,we determine the weights of the mutual influence of different stations at different times through an attention mechanism and extract spatiotemporal features through graph convolution on the RG and SG.Finally,we fused the spatiotemporal features and the original temporal features of stations for prediction.The comparison experiments on a railway bureau’s accurate railway passenger flow data show that the proposed F-SAGCN method improved the prediction accuracy and reduced the mean absolute percentage error(MAPE)of 46 stations to 7.93%.

Research on Railway Passenger Flow Prediction Method Based on GA Improved BP Neural Network

This paper chooses passenger flow data of some stations in China from January 2015 to March 2016, and the time series prediction model of BP neural network for railway passenger flow is established. But because of its slow convergence speed and easily falling into local optimal solution of the problem, we propose to improve the time series model of BP neural network by genetic algorithm to predict railway passenger flow. Experimental results show that the improved method has higher prediction accuracy and better nonlinear fitting ability.

Combination forecast for urban rail transit passenger flow based on fuzzy information granulation and CPSO-LS-SVM

In order to obtain the trend of urban rail transit traffic flow and grasp the fluctuation range of passenger flow better,this paper proposes a combined forecasting model of passenger flow fluctuation range based on fuzzy information granulation and least squares support vector machine(LS-SVM)optimized by chaos particle swarm optimization(CPSO).Due to the nonlinearity and fluctuation of the passenger flow,firstly,fuzzy information granulation is used to extract the valid data from the window according to the requirement.Secondly,CPSO that has strong global search ability is applied to optimize the parameters of the LS-SVM forecasting model.Finally,the combined model is used to forecast the fluctuation range of early peak passenger flow at Tiyu Xilu Station of Guangzhou Metro Line 3 in 2014,and the results are compared and analyzed with other models.Simulation results demonstrate that the combined forecasting model can effectively track the fluctuation of passenger flow,which provides an effective method for predicting the fluctuation range of short-term passenger flow in the future.

Hybrid Model for Short-Term Passenger Flow Prediction in Rail Transit

A precise and timely forecast of short-term rail transit passenger flow provides data support for traffic management and operation,assisting rail operators in efficiently allocating resources and timely relieving pressure on passenger safety and operation.First,the passenger flow sequence models in the study are broken down using VMD for noise reduction.The objective environment features are then added to the characteristic factors that affect the passenger flow.The target station serves as an additional spatial feature and is mined concurrently using the KNN algorithm.It is shown that the hybrid model VMD-CLSMT has a higher prediction accuracy,by setting BP,CNN,and LSTM reference experiments.All models’second order prediction effects are superior to their first order effects,showing that the residual network can significantly raise model prediction accuracy.Additionally,it confirms the efficacy of supplementary and objective environmental features.

基于熵值法的交旅融合服务区新增客流预测

[背景]路衍经济是行业转变发展方式、提升服务水平、增强发展韧性的重要选择。交旅融合产业是路衍经济的重要产业方向,高速公路服务区人流、车流汇集,具有天然的窗口平台优势和点状区位优势,是发展交旅融合产业的最佳场景。[目标]通过使用设施基本数据与历史交通流数据,采用熵值法对新建交旅融合服务区的新增客流进行预测,可以为服务区的建设和运营提供数据支持。[方法]以G75兰海高速公路太石服务区交旅融合项目为依托,基于承载力、价格等影响因素,构建了基于熵值法的吸引力模型,提出了交旅融合服务区新增客流的预测方法。采用熵值法确定模型系数,依据淡旺季各功能分区内设施的开放情况定量计算。[结果]客流计算结果显示,娱乐休闲区在开放期间吸引的客流最多,占到当月新增客流的61.1%。餐饮类与住宿类在淡旺季间的新增客流差距不大,而商务招待类的淡季客流显著高于旺季客流。[应用]通过计算交旅融合服务区中各功能区的新增客流量,可以为建成后的交旅融合服务区各功能区的开放计划及人员配备提供数据支撑,为提高服务区服务质量和服务满意度提供决策依据,以提高服务区的经营效益和资源利用率。

基于多源数据的福州地铁车站高峰客流预测研究

为提高地铁车站高峰客流预测精度,基于多源数据和统计学规律,分析了地铁车站高峰时段乘客进站特性,以此建立人岗比—高峰小时流量—超高峰小时系数三维模型和人口—岗位—高峰小时偏差系数三维模型。以福建省福州市目前已开通运营的站点为研究对象,分析其AFC数据,获取乘客进站和换乘信息,结合站点周边不同用地性质,对影响因素和关键参数进行研究,建立超高峰小时系数模型和高峰小时偏差系数模型。

基于3D卷积神经网络的城市轨道交通客流量预测方法

针对城市轨道交通的客流量预测问题,提出了一种基于3D卷积神经网络的客流量预测模型CNN3D⁃SP。将原始客流数据封装成张量块状结构作为网络的输入,有利于简化网络结构,提高了网络的可扩展性。再分别利用不同深度和不同尺寸的3D卷积核对输入数据进行时间维度和空间维度的特征提取,动态捕捉更深层次的线网客流特征。为了使特征提取网络对时间序列的突变部分更为敏感,在时间特征提取阶段引入ResNet残差网络结构和峰值注意力机制。然后将提取到的特征降维,利用2D卷积神经网络对线网客流数据进行拟合,得到目标时段多站点的客流预测值。本文选用南京地铁部分车站的客流量数据对模型进行验证。验证结果表明,相比现有的预测模型(ARIMA,SVR,LSTM),本文所提的CNN3D⁃SP模型具有更好的预测精度;在引入残差网络结构与峰值注意力之后,提高了网络对峰值客流的预测性能,预测结果的平均绝对误差较引入前降低了11.2%。

基于ASTLSTM的地铁乘客流量短时预测

地铁乘客流量预测是智能交通系统的重要环节,当前大多数预测模型较少对地铁乘客流量进行时空相关性建模,且未考虑空气质量等天气因素带来的影响,存在地铁乘客流量预测准确度不高的问题。针对以上问题,提出基于注意力机制的时空长短期记忆(ASTLSTM)网络的地铁乘客流量短时预测模型。首先,对数据进行预处理;然后,利用注意力机制与图卷积网络(GCN)、卷积神经网络(CNN)相融合,挖掘地铁数据中的时空相关性,并通过长短期记忆网络(LSTM)来提取空气质量数据中的外部特征;最后,通过特征融合得到地铁乘客流量预测结果。实验结果表明,ASTLSTM模型与LSTM、Conv LSTM等典型模型相比,在短期的地铁乘客流量预测上都有较高的准确度。

考虑延误特征的航站楼离港聚集客流预测方法

为满足航班延误下航站楼内资源规划与旅客管理对聚集客流预测所提出的高精度和高效率要求,本文提出一种融合延误特征的离港聚集客流预测方法。通过引入航班延误特征量化表征航站楼离港聚集客流的波动情况,探究航班延误下离港聚集客流波动规律和分布特征,构建基于自适应噪声完全集合经验模态分解(CEEMDAN)、排列熵算法(PE)以及鲸鱼优化算法(WOA)优化的长短期记忆神经网络(LSTM)的短期航站楼聚集客流预测模型。首先,应用CEEMDAN将聚集客流数据序列分解为若干模态分量(Intrinsic Mode Function, IMF)和残差量(Residual, Res),降低原序列中数据的复杂性和非平稳性影响;其次,为减小模型计算规模,同时提高预测效率和精度,采用PE算法对IMF分量进行熵值重构;最后,建立WOA-LSTM聚集客流预测模型,利用鲸鱼优化算法优化LSTM超参数,叠加重构分量的预测结果,得到最终的聚集客流预测值。将模型应用于长三角某枢纽机场进行实例验证。结果表明:CEEMDAN-PE-WOA-LSTM预测模型性能最优,相较单一的LSTM模型,候机大厅聚集客流预测的均方根误差、平均绝对误差以及百分比误差分别降低42.78%、44.00%及45.62%;相较CEEMDAN-WOA-LSTM模型,预测效率提高41.64%。本文所提模型能够有效拟合存在显著非线性和非平稳性特征的候机大厅聚集客流,具有较高的预测精度和运算效率。

基于出行链的铁路车站开通初期客流预测研究

【目的】为解决铁路车站开通初期客流预测缺乏数据支撑,以及估计不同交通方式成本来计算客流转移量结果与实际偏差较大的问题。【方法】以新开通南沿江城际铁路江阴站为例,考虑该线与既有高速及城际铁路的线位关系,基于城际出行链,依托旅客出行行为调查对区域内可选择的不同交通方式的城际出行效用进行定量计算,采用Multinominal Logit(MNL)模型,区分工作日与非工作日,结合客流分配现状对效用变量系数进行标定,据此预测南沿江城际铁路江阴站开通后发往南京、上海枢纽的客流量,同时结合城市经济和社会发展推演计算发往沿线其他站点的客流。【结果】结果表明,南沿江城际铁路江阴站的开通可以有效吸引江阴地区原本选择既有铁路、公路去往南京、上海的客流以及选择公路去往南沿江城际铁路沿线其他城市的客流。【结论】因此,南沿江城际铁路的开通能有效缓解辐射范围内既有铁路、公路的运输压力,客流预测结果还将支撑站点运力资源配置优化研究。

高速铁路运营客流量及收入集成估算系统研究

针对现行高速铁路决策阶段运营客流量和收入预测误差大,历史数据存在的线性、二维、来源不可靠等问题。在深入分析客流量和收入影响因素的非线性、非确定性、耦合、非平衡性等演变趋势和机理的基础上,深入挖掘历史数据,依据BPNN、改进FC、FIS、粗糙集等理论方法,建立高速铁路运营客流量和收入多维时空基础历史数据库,构建基于BIM、PYTHON的客流量和收入非线性多维智能预测系统,实现对高速铁路运营客流量和收入的准确预测和多维展现。筛选出高速铁路运营客流量和收入相关复杂性影响因素;构建高速铁路运营客流量和收入历史数据多维时空数据支持系统;构建BPNN、改进FC、FIS运营客流量和收入预测模型,提高方法拟合性和预测准确性、有效性;设计并创建基于BIM、PYTHON的高速铁路运营客流量和收入多维智能展现系统,为提升高速铁路投资决策可靠性、可视化、智能化提供了技术和方法支持。

基于可解释机器学习框架的列车乘车区段客流分布预测方法

为解释客运产品特征对列车乘车区段客流分布预测的影响,本文提出一种基于可解释机器学习框架的高速铁路列车乘车区段客流分布预测方法。首先,提出基于梯度提升树模型的高速铁路列车乘车区段客流分布预测框架,构建不同梯度提升树模型(GBDT、XGBoost、LightGBM及CatBoost)的高速铁路列车乘车区段客流分布预测模型;其次,计算特征贡献重要度,基于SHAP(SHapley Additive exPlanations)方法实现特征变量优化,揭示单一特征和交互特征与列车乘车区段客流分布预测的非线性关系。北京南—上海虹桥间列车客流分布预测结果表明:4种模型可精准预测客流分布结果,GBDT,XGBoost,LightGBM及CatBoost在测试集的决定系数分别为0.9664,0.9601,0.9680及0.9715;特征优化后,按贡献重要度排序依次为标杆车,票价,旅行时间,日期,星期,车次及出发时间;特征优化后,CatBoost-7模型在验证集中的决定系数为0.9458;日期和标杆车对客流分布预测呈现非线性正相关,旅行时间对客流分布预测呈现非线性负相关,低旅行时间、高票价及出发时间整点的标杆车对客流分布预测产生正向影响。本文研究结果能够为高速铁路客运产品设计提供一定参考价值。

多头图注意机制的深度学习地铁客流预测方法

挖掘地铁站点间的客流时空关联性是实现高精度地铁实时客流预测的关键。由于地铁网络中站点间客流的空间关联强弱性较为复杂且难以量化,从而导致客流预测过分依赖时间关联性。为克服该缺陷,提出一种多头图注意机制的深度学习地铁客流预测方法,通过构建面向地铁网络的多头图注意机制对多个关联站点间的客流空间关联性进行学习,得出差异化权重值,量化目标站点与关联站点群之间的客流空间关联强弱性;同时,将多头图注意机制融入长短时记忆学习模型,以量化后的空间关联性数据为输入,结合地铁客流量的时间关联性实现客流预测。实验结果表明,所提出的方法可行、有效,能够提升地铁客流预测精度且预测结果优于经典预测方法。

基于ARIMA与LSTM的铁路车站客流预测方法比较

精准的客流预测是车站客运组织优化的基础,是提高运营安全和运输效率的有效途径。以江门东站全年进站客流数据为研究对象,分别构建ARIMA时间序列模型与LSTM神经网络模型,从预测精度、计算速度、误差指标评价、模型适应性等方面分析比较两种预测模型对客流预测结果的差异性。结果表明,LSTM模型预测精度和拟合精确度更优,ARIMA模型计算速度更快。研究结果对客流预测方法选择有借鉴意义。

城市轨道交通初期运营客流预测技术研究

城市轨道交通新开通线路初期运营客流预测是开展初期运营安全评估和运营组织筹备的基础性工作。分析初期运营客流预测在预测期、研究对象、基础数据和预测模型方面的特点,总结该类预测工作开展的技术难点。在此基础上提出一套适用于初期运营客流预测的技术路线,阐述其中基础资料收集、交通调查、交通运行特征分析、出行需求预测、评估分析及敏感性分析的研究要点,并概括总结客流预测指标、对既有线的影响评估、需要重点关注的车站等研究内容。研究结果表明,初期运营客流预测重点在于准确把握城市及交通发展现状,而解读城市及交通发展现状资料并掌握其运行现状特点,研判其短期发展趋势是初期运营客流预测的基础性工作;针对宏观交通需求分析模型对微观客流指标预测的精度不足问题,可在城市宏观交通需求预测的基础上,通过开展各类交通调查建立城轨系统出行需求分析专题模型。本研究成果以期为初期运营客流预测的技术方法、研究内容等提供借鉴。

基于因果卷积和Informer模型的城市公交客流预测

为了有效缓解日常交通拥堵,倡导健康绿色出行,通过城市客流特征分析和预测精准掌握城市交通中的客流量分布特点,综合考虑天气、节假日等多种影响因素对客流量的影响,基于卷积神经网络提出了一种因果卷积自注意力的城市公交客流量CCN-Informer预测模型,改善输入序列局部信息提取能力,获取长期依赖信息,从而提高预测准确性。实验使用因果卷积自注意力进行依赖性特征提取,通过控制编码器的样本计算成本,采用生成式解码器提高模型预测效率。结果表明,CCN-Informer模型考虑周围重要特征因素影响,样本数据拟合程度明显提高,与基线模型对比,标准误差下降范围为4%~20%,有更高的预测精度和运行效率,验证了模型的有效性,可以为城市交通客流量预测提供依据。

基于Stacking集成学习的机场线短时客流预测研究

地铁机场线客流具有高度时变性,受机场航班影响使得精准的短时客流预测具有挑战性。综合考虑机场航班信息和机场线路历史客流,构建了一种以随机森林(RF)、LightGBM(light gradient boosting machine)、梯度提升决策树(GBDT)和逻辑回归算法作为集成学习器,基于叠加(Stacking)集成模型的机场线路短时客流预测模型。以北京地铁大兴机场线为实例进行验证,并与Informer和长短时记忆神经网络(long short-term memory,LSTM)两种基线模型进行对比。结果表明,考虑航班信息和机场线历史客流的双通道预测效果明显优于仅考虑机场线历史客流的单通道预测;Stacking模型在各项指标中均表现出优越的性能,其中,在96步长(24 h)下的预测效果最好,预测进站客流的平均绝对误差为7.66,预测出站客流的平均绝对误差为4.67;分析航班信息特征对预测模型的影响,发现离港航班信息重要性不如到港航班,这与离港旅客提前到达机场时间差异较大有关。

域自适应动态图卷积网络下的地铁客流预测

针对客流预测中存在因数据量有限导致模型训练过程中出现高方差和泛化性差等问题,提出一种域自适应动态图卷积网络(GCN-DANN)。通过构建地铁线路的节点网络拓扑结构,并利用动态图卷积网络提取相邻站点之间的流量、站点所属线路的交通负载以及不同线路之间的流量传播等关联特征。同时采用迁移学习自适应对齐源域和目标域的特征,减少因数据分布不一致而导致预测性能低等现象。最后,通过全连接层将源域和目标域中的特征进行信息融合,进而弥补训练过程出现高方差和泛化性差等缺陷。在深圳地铁数据集上对模型训练,分别在杭州地铁全样本和20%样本数据集上进行测试和验证。实验结果表明,在20%样本数据集下,GCN-DANN网络与经典预测网络相比,MAE、RMSER和MAPE分别平均下降5.34%、6.07%、2.97%。在全样本数据集下,GCN-DANN在20%样本基础上的三项指标分别下降2.76%、1.77%、3.5%,相较于其他经典网络下降幅度最小。研究可解决实际应用中因数据稀缺导致预测效果差的问题。